Configuration

MSAlign score fusion +Filter (overbar) [MSAlign v2]

MSAlign retrieves candidate molecules from tandem mass spectra by aligning pretrained molecular and spectral representations.

Sources (2)

https://arxiv.org/abs/2605.19752: page.html; msalign: Primary paper PDF · MSAlign paper Section 3 Architecture and Training, Section 4 on splitting, and Section 5.1 Experimental Setting

3 evaluations · 9 results

How it worksMSAlign workflow

Conceptual summary of the documented data flow; optional inputs and configured downstream stages must be reported for a reproducible evaluation.

Sources (2)

https://arxiv.org/abs/2605.19752: page.html; msalign: Primary paper PDF · MSAlign paper Section 3 Architecture and Training, Section 4 on splitting, and Section 5.1 Experimental Setting

Overview

Model type

Frozen spectral/molecular encoders with learned alignment projections

Inputs

MS/MS spectrum and a set of candidate molecular structures.

Outputs

Candidate-molecule retrieval scores in a shared representation space.

Access

Official project documentation and implementation: https://arxiv.org/abs/2605.19752

Sources (2)

https://arxiv.org/abs/2605.19752: page.html; msalign: Primary paper PDF · MSAlign paper Section 3 Architecture and Training, Section 4 on splitting, and Section 5.1 Experimental Setting

limited source coverage · Automated source review, 2026-09-16. All specifications and missing details

Evaluations and results

3 evaluations · 9 results. Different protocols are not a single leaderboard.

Filter evaluations

Tested model or methodProtocolDatasetEvaluation setupMetricEvidence origin

Applied filters: All linked evaluations

Exact evaluated configurations and original reported results
Tested configuration	Protocol and dataset	Finding	Evidence and details
Configuration: MSAlign score fusion +Filter (overbar) [MSAlign v2]	Protocol: MSAlign v2 retrieval: massspecgym-formula-formula-oracle Dataset subset: MSAlign v2 massspecgym-formula-formula-oracle	60.4 ± 3.7% recall_at_1 percent · higher Uncertainty: type: standard_deviation; reported spread: 3.7; scope: Three random splitting seeds. Coverage: Not reported scored / Not reported eligible	Author-reported evaluation · Source checked Methods, coverage and source MSAlign score fusion +Filter (overbar) massspecgym-formula-formula-oracle Not reported Aggregation: Not reported MSAlign arXiv v2 Table 3 · Table 3 p7; massspecgym-formula block, R@1 row, MSAlign score fusion +Filter (overbar) column
Configuration: MSAlign score fusion +Filter (overbar) [MSAlign v2]	Protocol: MSAlign v2 retrieval: massspecgym-formula-formula-oracle Dataset subset: MSAlign v2 massspecgym-formula-formula-oracle	93.6 ± 1.1% recall_at_20 percent · higher Uncertainty: type: standard_deviation; reported spread: 1.1; scope: Three random splitting seeds. Coverage: Not reported scored / Not reported eligible	Author-reported evaluation · Source checked Methods, coverage and source MSAlign score fusion +Filter (overbar) massspecgym-formula-formula-oracle Not reported Aggregation: Not reported MSAlign arXiv v2 Table 3 · Table 3 p7; massspecgym-formula block, R@20 row, MSAlign score fusion +Filter (overbar) column
Configuration: MSAlign score fusion +Filter (overbar) [MSAlign v2]	Protocol: MSAlign v2 retrieval: massspecgym-formula-formula-oracle Dataset subset: MSAlign v2 massspecgym-formula-formula-oracle	82.1 ± 3.3% recall_at_5 percent · higher Uncertainty: type: standard_deviation; reported spread: 3.3; scope: Three random splitting seeds. Coverage: Not reported scored / Not reported eligible	Author-reported evaluation · Source checked Methods, coverage and source MSAlign score fusion +Filter (overbar) massspecgym-formula-formula-oracle Not reported Aggregation: Not reported MSAlign arXiv v2 Table 3 · Table 3 p7; massspecgym-formula block, R@5 row, MSAlign score fusion +Filter (overbar) column
Configuration: MSAlign score fusion +Filter (overbar) [MSAlign v2]	Protocol: MSAlign v2 retrieval: massspecgym-mces-formula-oracle Dataset subset: MSAlign v2 massspecgym-mces-formula-oracle	37.7 ± 1.3% recall_at_1 percent · higher Uncertainty: type: standard_deviation; reported spread: 1.3; scope: Two MCES validation/test-swapped variants per methods, despite blanket three-split table caption. Coverage: Not reported scored / Not reported eligible	Author-reported evaluation · Source checked Methods, coverage and source MSAlign score fusion +Filter (overbar) massspecgym-mces-formula-oracle Not reported Aggregation: Not reported MSAlign arXiv v2 Table 3 · Table 3 p7; massspecgym-mces block, R@1 row, MSAlign score fusion +Filter (overbar) column
Configuration: MSAlign score fusion +Filter (overbar) [MSAlign v2]	Protocol: MSAlign v2 retrieval: massspecgym-mces-formula-oracle Dataset subset: MSAlign v2 massspecgym-mces-formula-oracle	87.1 ± 2.3% recall_at_20 percent · higher Uncertainty: type: standard_deviation; reported spread: 2.3; scope: Two MCES validation/test-swapped variants per methods, despite blanket three-split table caption. Coverage: Not reported scored / Not reported eligible	Author-reported evaluation · Source checked Methods, coverage and source MSAlign score fusion +Filter (overbar) massspecgym-mces-formula-oracle Not reported Aggregation: Not reported MSAlign arXiv v2 Table 3 · Table 3 p7; massspecgym-mces block, R@20 row, MSAlign score fusion +Filter (overbar) column
Configuration: MSAlign score fusion +Filter (overbar) [MSAlign v2]	Protocol: MSAlign v2 retrieval: massspecgym-mces-formula-oracle Dataset subset: MSAlign v2 massspecgym-mces-formula-oracle	67.2 ± 1.0% recall_at_5 percent · higher Uncertainty: type: standard_deviation; reported spread: 1.0; scope: Two MCES validation/test-swapped variants per methods, despite blanket three-split table caption. Coverage: Not reported scored / Not reported eligible	Author-reported evaluation · Source checked Methods, coverage and source MSAlign score fusion +Filter (overbar) massspecgym-mces-formula-oracle Not reported Aggregation: Not reported MSAlign arXiv v2 Table 3 · Table 3 p7; massspecgym-mces block, R@5 row, MSAlign score fusion +Filter (overbar) column
Configuration: MSAlign score fusion +Filter (overbar) [MSAlign v2]	Protocol: MSAlign v2 retrieval: spectraverse-formula-oracle Dataset subset: MSAlign v2 spectraverse-formula-oracle	44.4 ± 2.6% recall_at_1 percent · higher Uncertainty: type: standard_deviation; reported spread: 2.6; scope: Three random splitting seeds. Coverage: Not reported scored / Not reported eligible	Author-reported evaluation · Source checked Methods, coverage and source MSAlign score fusion +Filter (overbar) spectraverse-formula-oracle Not reported Aggregation: Not reported MSAlign arXiv v2 Table 3 · Table 3 p7; spectraverse block, R@1 row, MSAlign score fusion +Filter (overbar) column
Configuration: MSAlign score fusion +Filter (overbar) [MSAlign v2]	Protocol: MSAlign v2 retrieval: spectraverse-formula-oracle Dataset subset: MSAlign v2 spectraverse-formula-oracle	87.5 ± 0.9% recall_at_20 percent · higher Uncertainty: type: standard_deviation; reported spread: 0.9; scope: Three random splitting seeds. Coverage: Not reported scored / Not reported eligible	Author-reported evaluation · Source checked Methods, coverage and source MSAlign score fusion +Filter (overbar) spectraverse-formula-oracle Not reported Aggregation: Not reported MSAlign arXiv v2 Table 3 · Table 3 p7; spectraverse block, R@20 row, MSAlign score fusion +Filter (overbar) column
Configuration: MSAlign score fusion +Filter (overbar) [MSAlign v2]	Protocol: MSAlign v2 retrieval: spectraverse-formula-oracle Dataset subset: MSAlign v2 spectraverse-formula-oracle	71.4 ± 1.5% recall_at_5 percent · higher Uncertainty: type: standard_deviation; reported spread: 1.5; scope: Three random splitting seeds. Coverage: Not reported scored / Not reported eligible	Author-reported evaluation · Source checked Methods, coverage and source MSAlign score fusion +Filter (overbar) spectraverse-formula-oracle Not reported Aggregation: Not reported MSAlign arXiv v2 Table 3 · Table 3 p7; spectraverse block, R@5 row, MSAlign score fusion +Filter (overbar) column

Source checking is not independent reproduction. Release 2026-09-30-e37e3ab1284d.

Use this model

How it works, versions and access

Related profile: MSAlign. This page retains the exact record and its evaluation context.

This configuration

Exact Table 3 source implementation. Requires target molecular formula oracle.

record: MSAlign score fusion +Filter (overbar) [MSAlign v2]
configuration: Not reported
entity type: Configuration

How it works

MSAlign retrieves candidate molecules from tandem mass spectra by aligning pretrained molecular and spectral representations. Frozen DreaMS and ChemBERTa encoders connected by lightweight MLP projections trained with a candidate-based contrastive objective. The documented inputs are MS/MS spectrum and a set of candidate molecular structures. The output consists of candidate-molecule retrieval scores in a shared representation space.

Sources (2)

https://arxiv.org/abs/2605.19752: page.html; msalign: Primary paper PDF · MSAlign paper Section 3 Architecture and Training, Section 4 on splitting, and Section 5.1 Experimental Setting

Versions and reproducibility

MSAlign arXiv:2605.19752v1, submitted 19 May 2026. The applicable input limits require configuration-specific checking.

Sources (2)

https://arxiv.org/abs/2605.19752: page.html; msalign: Primary paper PDF · MSAlign paper Section 3 Architecture and Training, Section 4 on splitting, and Section 5.1 Experimental Setting

Strengths, limitations and unresolved questions

Strengths and limitations

Strengths and considerations

Keeps the large encoders frozen while learning alignment projections for the retrieval task.
Sources (2)
https://arxiv.org/abs/2605.19752: page.html; msalign: Primary paper PDF · MSAlign paper Section 3 Architecture and Training, Section 4 on splitting, and Section 5.1 Experimental Setting

Limitations and conditions

Candidate construction and splitting change the retrieval problem. The authors explicitly discuss the trade-off between leakage control and distribution shift.
Sources (2)
https://arxiv.org/abs/2605.19752: page.html; msalign: Primary paper PDF · MSAlign paper Section 3 Architecture and Training, Section 4 on splitting, and Section 5.1 Experimental Setting

Profile review details

Inspected pinned official documentation, relevant implementation files and named primary-paper sections. Claims are limited to those artifacts. Remaining field extraction and identity conflicts are explicit; no new performance claims, model runs or human review are implied.

Stable record: discovery-model-msalign

Specifications

Inputs, training, access and other details

Explanatory profile: limited source coverage · Automated source review, 2026-09-16. Review applies to the cited claims; unresolved fields are listed below. Numerical results retain their own review status.

Inputs, outputs and configuration
Property	Description and evidence
Model type	Frozen spectral/molecular encoders with learned alignment projections Sources (2) https://arxiv.org/abs/2605.19752: page.html; msalign: Primary paper PDF · MSAlign paper Section 3 Architecture and Training, Section 4 on splitting, and Section 5.1 Experimental Setting
Architecture	Frozen DreaMS and ChemBERTa encoders connected by lightweight MLP projections trained with a candidate-based contrastive objective. Sources (2) https://arxiv.org/abs/2605.19752: page.html; msalign: Primary paper PDF · MSAlign paper Section 3 Architecture and Training, Section 4 on splitting, and Section 5.1 Experimental Setting
Inputs	MS/MS spectrum and a set of candidate molecular structures. Sources (2) https://arxiv.org/abs/2605.19752: page.html; msalign: Primary paper PDF · MSAlign paper Section 3 Architecture and Training, Section 4 on splitting, and Section 5.1 Experimental Setting
Outputs	Candidate-molecule retrieval scores in a shared representation space. Sources (2) https://arxiv.org/abs/2605.19752: page.html; msalign: Primary paper PDF · MSAlign paper Section 3 Architecture and Training, Section 4 on splitting, and Section 5.1 Experimental Setting
Parameters	Approximately 4M trainable projection parameters; frozen DreaMS and ChemBERTa backbones are reported as 96M and 92M respectively. Sources (2) https://arxiv.org/abs/2605.19752: page.html; msalign: Primary paper PDF · MSAlign paper Section 3 Architecture and Training, Section 4 on splitting, and Section 5.1 Experimental Setting
Known versions	MSAlign arXiv:2605.19752v1, submitted 19 May 2026. Sources (2) https://arxiv.org/abs/2605.19752: page.html; msalign: Primary paper PDF · MSAlign paper Section 3 Architecture and Training, Section 4 on splitting, and Section 5.1 Experimental Setting
Training data	Projection layers are fitted separately on the NPLIB1, MassSpecGym or Spectraverse training splits. The paper distinguishes spectrum/molecule pair counts from unique molecules and controls candidate retrieval using mass matching. Sources (2) https://arxiv.org/abs/2605.19752: page.html; msalign: Primary paper PDF · MSAlign paper Section 3 Architecture and Training, Section 4 on splitting, and Section 5.1 Experimental Setting
Training cutoff	NPLIB1, MassSpecGym and Spectraverse are separately split training resources. The inspected paper does not define one latest measurement date covering all three. · Not reported in inspected sources Sources (2) https://arxiv.org/abs/2605.19752: page.html; msalign: Primary paper PDF · MSAlign paper Section 3 Architecture and Training, Section 4 on splitting, and Section 5.1 Experimental Setting
Context limits	The pipeline inherits spectrum and molecule preprocessing from its frozen DreaMS and ChemBERTa encoders. The inspected MSAlign architecture section does not state a single joint input limit. · Not reported in inspected sources Sources (2) https://arxiv.org/abs/2605.19752: page.html; msalign: Primary paper PDF · MSAlign paper Section 3 Architecture and Training, Section 4 on splitting, and Section 5.1 Experimental Setting
Weights licence	The inspected preprint does not state distribution terms for the learned projection checkpoints. Frozen encoder licences remain separate from projection-weight rights. · Not reported in inspected sources Sources (2) https://arxiv.org/abs/2605.19752: page.html; msalign: Primary paper PDF · MSAlign paper Section 3 Architecture and Training, Section 4 on splitting, and Section 5.1 Experimental Setting; page.html: inspected official source
Access	Official project documentation and implementation: https://arxiv.org/abs/2605.19752 Sources (2) https://arxiv.org/abs/2605.19752: page.html; msalign: Primary paper PDF · MSAlign paper Section 3 Architecture and Training, Section 4 on splitting, and Section 5.1 Experimental Setting
Code licence	The inspected preprint describes the algorithm but does not supply a separate code licence; no code-distribution permission is inferred from the paper licence. · Not reported in inspected sources Sources https://arxiv.org/abs/2605.19752: page.html · page.html: inspected official source

Evidence

Source checking verifies the cited claim or transcription. It does not establish independent reproduction.

Evidence table

Inspect claims, sources and review details

Trace each statement to its source and review. A context-only reference supports the record generally; it does not verify an individual field. Source checking does not reproduce an experiment.

One row per statement and cited source. Multiple citations are not independent evaluations. Shared locators are labelled explicitly.

1 evidence row matching the loaded filters

Claims, original sources and review scope · Release 2026-09-30-e37e3ab1284d
Property and statement	Original source and location	Review and provenance
Relationship: variant of discovery-model-msalign Individual claims	MSAlign arXiv v2 Table 3 Original source ↗ Table 3 source column MSAlign score fusion +Filter (overbar); Section 5.1 Baselines Version: 2605.19752v2, 25 September 2026 Retrieved: 2026-09-30	source checked automated source review · 2026-09-30T21:40:59.576893+00:00 Audit details Field: `links:variant_of:discovery-model-msalign` Claim: uc20260930-config-family-df1940bc6b4c4a7e Source artifact SHA-256: `a31d1167b0ded72f64717881dd6cba0946a56b495d76f169d8e44115c3266fda` Hash scope: Hash scope not separately documented; inspect source record Inspected artifact

All evidence (CSV)All evidence (JSONL)How to read the evidence

Sources and history

View linked audit checks and correction history

Release 2026-09-30-e37e3ab1284d · Record review: source checked

1 source records and release history

MSAlign arXiv v2 Table 3 · Original source · 2605.19752v2, 25 September 2026

Download this release

Technical metadata and extraction receipts

Stable ID: uc20260930-msalign-v2-config-msalign-score-fusion-filter-overbar

areas: metabolomics
source locator: Table 3 column MSAlign score fusion +Filter (overbar); Sections 3 and 5.1
missing metadata: checkpoint hash: unextracted
review: method: automated_source_review; reviewer: Codex experimental-use-case curation; date: 2026-09-30; reviewed at: 2026-09-30T21:40:59.576893+00:00; scope: Primary-source transcription and catalogue applicability audit; not human scientific review or independent experimental reproduction.
reported name: MSAlign score fusion +Filter (overbar)

Related records

variant of: MSAlign
subject: MSAlign score fusion +Filter (overbar) [MSAlign v2] family identity
configuration: MSAlign score fusion +Filter (overbar) massspecgym-formula-formula-oracle
configuration: MSAlign score fusion +Filter (overbar) massspecgym-mces-formula-oracle
configuration: MSAlign score fusion +Filter (overbar) spectraverse-formula-oracle

Suggest a correction